Beyond the Bubble

The Weekender is my bi-weekly take on macro shifts and emerging themes. It’s not investment advice — or even our firm’s official view. I aim simply to inform, challenge, and maybe entertain. If you’d like this in your inbox every other Saturday morning via Northern Trust, click the subscribe to insights button in the lower left corner and select The Weekender.

Rise of the Machines

Once Elon Musk cracks the hand (achieving human-level dexterity and degrees of freedom), vertically integrates the supply chain, and scales the factory, then humanoid robotics could become the largest application of machine intelligence. Possibly the largest market, period.

Its size? Consider Apple’s 1.5 billion users: the wealthiest consumer cohort on Earth with a taste for tech. If each household ends up with one robot, that’s 1.5 billion units. Each will require at least one AI chip. The problem? The world currently produces about 6 million AI chips annually, mostly in Taiwan. That might explain the strategic stake-building in Intel — and the soaring demand for computational power. See Oracle’s sales backlog for clues.

Strategic Decoupling

Intel’s moves and China’s pivot toward chip self-sufficiency raise a question: does this shift alter Taiwan’s strategic calculus? If China and the U.S. reduce reliance on Taiwanese fabs, could Taiwan’s geopolitical significance — and the risk of conflict — diminish? Worth considering, especially for those who labelled China “un-investable” based on this very risk.

China, Evolving Bull

Weekender readers know our China thesis: from “deep value” in February 2023 to a secular bull market today. Risks remain — geopolitics, trade — but the macro pivot is clear: from credit-heavy, property-centric growth to equity-led capital formation. This transformation — “equitization” — is being driven by coordinated monetary, fiscal and administrative levers. The goal? Convert household savings (at record % of gross domestic product) into enterprise capital. Reach out to us if you want more detail.

A Cautionary Tale

In the 1940s, one of the largest unions in the urban U.S. was elevator operators. Post-World War II strikes in New York paralyzed business and accelerated adoption of automated lifts — tech that had existed for decades but needed a catalyst. By the 1950s, the union was gone. A cautionary tale for certain strikes I endured last week in London. We have self-driving cars on roads, but not trains on rails? London Tube drivers, are you listening?

Chips, Power and the Oracle of Scale

“The Oracle deal is one of the largest cloud contracts ever signed. It will require 4.5 gigawatts of power — more than two Hoover Dams or enough to power four million homes.” WSJ. As we noted last time, U.S. President Donald Trump has said “if we want to win the AI race, we need to win the electricity race first”. Another reason to expect a broadening rally. The recent rotation toward industrials and tangible assets suggests that the picks and shovels of the ethereal world (like Nvidia) may extend to the material world. Perhaps that’s what the S&P Metals and Mining Select Industries Index is signalling. It’s up 50% year-to-date — a record high.

Cheaper Hedging

Fed rate cuts often trigger a reflexive “buy stocks” response during economic recovery. They also reduce the cost of shorting the dollar, making it a less punitive carry trade and more attractive from an overlay perspective. As a result, global investors may increase outright dollar hedging. Importantly, hedging decisions are separate from underlying asset exposure. Indeed the most recent data shows investors are increasing, not exiting, their dollar-denominated holdings. All the while the dollar continues to drift lower. That’s good news for earnings.

The Reluctant Rally

Markets seem to be warming to U.S. Treasury Secretary Scott Bessent’s worldview: a lower neutral rate and a Fed willing to run things hot into a tech-led, non-inflationary growth cycle—think Greenspan in the '90s — making this chart below more prescient (and all the more painful for those reluctant to participate). One thing to note, when the Federal Reserve cuts with markets at all-time-highs, and the economy isn’t on the brink (think September 2007), stocks have been up 12 months later — 100% of the time since 1990. Will history rhyme? See Exhibit 1.

EXHIBIT 1: 2020S VS. 1990S: A REPEAT?

Robotic Money

On September 8, Nasdaq filed to list tokenized stocks and ETFs on its main exchange—potentially enabling 24/5 trading by the third quarter of 2026. Days later, BlackRock signalled similar plans. Weeks earlier, Galaxy Digital announced it will tokenize its shares on Solana—the first U.S. public company to put equity on a major blockchain. Helping drive this shift are stablecoins, enabling instant settlement and digital collateral as the bridge from TradFi to DeFi. But their role doesn’t stop there. This week, Google launched the Agents Payments Protocol (AP2), allowing AI agents not only to communicate — but to transact. Stablecoins may help make that possible, creating a likely positive correlation between agentic commerce and stablecoin adoption. That’s bullish for infrastructure layers like Ethereum, stablecoin AUM, and — by extension — demand for Treasury bills.

Pets.com PTSD

I remember the 2000 dot.com bubble well. My employer underwrote Pets.com — days before the Nasdaq Composite Index peak. I still bear the scars (and the PTSD). So when asked to compare today’s market to then, I do so through gritted teeth.
But there are differences…

Then vs. Now: Bubble Math

Yes, tech concentration is high — but unlike 2000, earnings concentration matches it: tech drives over 50% of S&P 500 Index earnings-per-share growth. Back then, profits were scarce (cue “price to eyeballs”); today, capex is funded by cash flow, not frothy capital markets. Back then, technology credit spreads widened before the equity peak; today, they’re near cycle lows. Back then, investors sold business plans; now, established franchises carry backlogs 6 times sales (see Oracle). Back then, the Fed was hiking; now, it’s cutting. And leaders are cheaper on earnings: Oracle has a two year forward price-to-earnings ratio of 29 now versus about 80 then; Microsoft is about 23 now versus 53 then — with higher margins and stronger cash generation. So yes, prices are rich — but not irrational relative to earnings power, which ultimately rules. It could stall, of course, but more likely broadens before it breaks.

Compounding: The Real Flywheel

But perhaps the biggest difference between now and then?Time.

Time amplifies compounding.

Think Buffett — 94% of his wealth came after age 60. Or the 42nd fold of a paper would create enough thickness to reach the moon.

Consider, the ChatGPT moment for the dot.com era was, I believe, Netscape, which launched early 1995.

Five years later, the bubble burst.

Not much time to build-out the base. To generate a flywheel. To compound.

Now? We’re building off a bigger base — of data, computers, infrastructure. AI’s growth is not linear, it’s exponential. Each breakthrough amplifies the next. Like when a Tesla sees a kangaroo for the first time — within minutes, every Tesla knows how it moves. That’s why former Google CEO Eric Schmidt says AI is underhyped. Why Nvidia CEO Jensen Huang sees exponential growth. Why Oracle CEO Safra Catz calls it insatiable. Why Palantir sees a 10-fold growth in sales. And Anthropic says computational power doubles every six months.

And this is before computers train computers.

So, expensive? Yes. A short-term top? Maybe.

But a bubble?

Not yet.

They look more like Exhibit 2.

EXHIBIT 2: CHATGPT LIKE NETSCAPE?

Have a great weekend.

Gary